The role of relaxin-3/RXFP3 signalling in anxiety, stress and feeding in the mouse

© 2017 Dr. Cary ZhangRelaxin-3 is a neuropeptide expressed by neurons of the brainstem nucleus incertus (NI). Relaxin-3 signalling via its cognate receptor, RXFP3, is proposed to be involved in the neuromodulatory control of arousal- and stress-related neural circuits, based largely on experimental studies completed in the rat. For example, central administration studies revealed that native relaxin-3 increased hypothalamic-pituitary-adrenal (HPA) axis activity, and a relaxin-3 analogue selective for RXFP3 relative to RXFP1 reduced anxiety-like behaviours and promoted feeding. However, far less evidence has been collected for similar roles in the mouse and other major experimental species. The aim of these studies was to extend these previous findings and further explore the ability of the relaxin-3/RXFP3 system to modulate: (i) anxiety-like behaviour; (ii) HPA axis activity, reflected by corticosterone levels and activity of paraventricular hypothalamic nucleus neurons, and; (iii) feeding behaviour. These studies were conducted in adult C57BL/6J mice, to provide an important species comparison, and also employed several relevant transgenic mouse strains. Studies examining effects of RXFP3 modulation on anxiety-like behaviours were conducted using established behavioural tests - the elevated plus maze (EPM), light/dark (L/D) box, large open field (LOF) and social interactions tests. These studies revealed that intracerebroventricular (icv) injection of the RXFP3 agonist, RXFP3-A2 (R3(B)/R3(A11-24)), did not affect baseline anxiety levels in mice in the EPM, L/D box and LOF. Subsequently, effects of RXFP3-A2 on increased levels of anxiety induced by FG-7142 (an inverse benzodiazepine agonist) were examined, to determine whether increased RXFP3 signalling was able to reduce elevated anxiety. In the EPM, FG-7142 treatment did not increase anxiety relative to control levels, and consistent with previous observations, RXFP3-A2 did not affect anxiety-like behaviour at these ‘baseline’ levels. However, in the L/D box, FG-7142 effectively increased anxiety-like behaviour, which was restored to control levels by RXFP3-A2, suggesting an anxiolytic effect of central RXFP3 activation. The role of endogenous relaxin-3/RXFP3 signalling was assessed by examining anxiety-like behaviours after icv administration of the RXFP3 antagonist, R3(B1-22)R. RXFP3 inhibition resulted in a significant increase in anxiety-like behaviour in the EPM, but not the L/D box or LOF. These data suggest endogenous relaxin-3 signalling plays a subtle or context-specific role in modulation of basal innate anxiety in mice. Previous studies in rats led to the proposal that relaxin-3/RXFP3 signalling is associated with activation of the HPA axis and corticotropin-releasing hormone (CRH) neurons of the PVN, but direct evidence for RXFP3-related actions on the activity of hypothalamic CRH neurons was not provided in these earlier reports. In these studies, the characteristics of the relaxin-3/RXFP3 system in mouse hypothalamus were investigated, with the aim of determining the precise nature of relaxin-3/RXFP3 interactions with HPA axis activity in the mouse. Despite the reported ability of relaxin-3 administration to increase plasma corticosterone levels in rats, administration of RXFP3-A2 icv or directly into the PVN of C57BL/6J mice did not alter serum corticosterone. Similarly, there were no differences between serum corticosterone levels in RXFP3 KO and wildtype mice at baseline and after an acute restraint stress, despite detection of the predicted stress-induced increase in serum corticosterone. I examined the nature of the relaxin-3 innervation of PVN in wildtype mice and in CRH-IRES-Cre;Ai14 mice that co-express the tdTomato fluorophore in CRH neurons, and identified abundant relaxin-3 fibres in the peri-PVN region, but only sparse fibres associated with the soma of the densely-packed CRH neurons, suggesting a lack of influence of relaxin-3 on these neurons, or at best, a mild modulatory impact. In whole-cell voltage-clamp recordings of tdTomato-positive CRH neurons in these mice, despite a lack of any apparent direct effects on their activity, a decrease in spontaneous excitatory post-synaptic currents (sEPSCs) was observed following local application of RXFP3-A2, consistent with an activation of inhibitory RXFP3 on presynaptic glutamatergic afferents in the PVN region. These data provide evidence that the effect of RXFP3 activation on PVN CRH neurons is indirect, which should potentially result in a decrease in HPA axis activity rather than an increase, as observed in rats. Further studies using more appropriate assays will be required to determine if this circuit is relevant in vivo. RXFP3 modulation has also been found to alter feeding in adult rats and independent experimental evidence demonstrates that central RXFP3 activation by relaxin-3 or RXFP3-A2 increases feeding behaviour in rats. Therefore, in the current studies, the effects of RXFP3 modulation on feeding and associated behaviours were examined in mice. Following direct administration of RXFP3-A2 into the PVN, no significant difference in food consumption relative to vehicle controls was observed. However, icv injection of an RXFP3 antagonist in C57BL/6J mice was able to significantly reduce: food anticipatory activity, food consumption during the early dark phase, food consumption driven by mild food deprivation, and consumption of palatable food. In subsequent studies, a new RXFP3 antagonist, KKK-pK-R3(B6-22)R (KKK-R), which was chemically modified to promote improved plasma stability and blood-brain-barrier permeability, was tested to determine if an intraperitoneal (ip) injection was able to reduce food consumption in mice. Although peripheral administration of KKK-R was able to reduce palatable food consumption and regular chow consumption during the early dark phase relative to saline-injected controls, a similar effect was observed in RXFP3 KO mice, suggesting non-specific ‘off target’ effects of the modified peptide at the doses used. However, the replacement of the key C terminal arginine residue within KKK-R with glutamic acid dramatically reduced its pharmacological affinity for RXFP3, and this modified compound displayed no effect on feeding at the same dose, suggesting KKK-R may target non-RXFP3 receptors due to the increased affinity generated by the C terminus, arginine. These studies have increased our knowledge of the role played by relaxin-3/RXFP3 signalling in modulating anxiety-like behaviour, the HPA axis and PVN, and feeding behaviour in mice. Furthermore, these studies have revealed species differences between mice and rats, which may have implications for predicting how conserved various features of relaxin-3/RXFP3 biology might be in humans, and these differences need to be taken into account when conducting and comparing further studies in these two major experimental species. Altogether, the findings presented in this thesis suggest that the relaxin-3/RXFP3 system is a potential therapeutic target for agents designed to treat feeding and anxiety disorders. Present findings also provide a platform for continued studies to achieve this long term goal